1. Proposed Text for DL subcarrier permutation and UL tile permutation IEEE 802.16 Presentation Submission Template (Rev. 9) Document Number: IEEE C802.16m-09/0582r5 Date Submitted: 2009-03-09 Source: Taeyoung Kim, Jeongho Park, Kichun Cho, Jaeweon Cho, Voice:+82-31-279-0202 Hokyu Choi, Heewon Kang E-mail:ty33.kim@samsung.com Samsung Electronics Co., Ltd. 416 Maetan-3, Suwon, 443-770, Korea Jong-Kae (JK) Fwu, Minh-Anh Vuong, Huaning Niu, Rongzhen Yang, Yuval Lomnitz, Wei Guan, Sassan Ahmadi, Hujun Yin E-mail:jong-kae.fwu@intel.com Intel Corporation Jihyung Kim, Wooram Shin, Dong Seung Kwon E-mail:savant21@etri.re.kr ETRI Yu-Tao Hsieh, Pang-An Ting, Zheng Yan-Xiu E-mail: yutaohsieh@gmail.com ITRI

2. Venue: IEEE 802.16m Session#60, Vancouver, Canada IEEE 802.16m-09/0012, “Call for Comments on Amendment Working Document”. Base Contribution: None Purpose: Discussion and Approval Notice: This document does not represent the agreed views of the IEEE 802.16 Working Group or any of its subgroups. It represents only the views of the participants listed in the “Source(s)” field above. It is offered as a basis for discussion. It is not binding on the contributor(s), who reserve(s) the right to add, amend or withdraw material contained herein. Release: The contributor grants a free, irrevocable license to the IEEE to incorporate material contained in this contribution, and any modifications thereof, in the creation of an IE EE Standards publication; to copyright in the IEEE’s name any IEEE Standards publication even though it may include portions of this contribution; and at the IEEE’s sol e discretion to permit others to reproduce in whole or in part the resulting IEEE Standards publication. The contributor also acknowledges and accepts that this contributio n may be made public by IEEE 802.16. Patent Policy: The contributor is familiar with the IEEE-SA Patent Policy and Procedures: and.http://standards.ieee.org/guides/bylaws/sect6-7.html#6http://standards.ieee.org/guides/opman/sect6.html#6.3 Further information is located at and.http://standards.ieee.org/board/pat/pat-material.htmlhttp://standards.ieee.org/board/pat

3. Motivation In the current IEEE 802.16m Amendment Working Document (IEEE80216m-09/0010), –Subcarrier permutation for DL is NOT determined yet –Tile permutation for UL is NOT determined yet This contribution shows the evaluation results to compare the permutation rules proposed from many companies (e.g. Intel, LGE, Samsung) –For Downlink, LLS results under the multicell environment –For Uplink, SLS and Hitting Count Results 3/14

4. Uplink Tile Permutation Part 1 4/17

5. Issue Different Permutation Approaches –Intel/Samsung : –LG : 5/17 where and Tile(s,n,t) = L DRU,FPi  n + g(PermSeq(), s, n, t) PermSeq() is permutation sequence generated with SEED={IDcell*1367} mod 2 10. g(PermSeq(),s,n, t) = {PermSeq[(n+107*s+t) mod L DRU,FPi ]+UL_PermBase} mod L DRU,FPi, UL_PermBase is set to IDcell.

6. Hitting Count Verification Methodology –Select 2 Sector IDs from total 768 Total number of cases : combination 2 among 768 = 294528 cases –Count the number of tiles two sectors share Full loading case  obviously 100% hitting Not full loading case  worst case is 100% hitting –Results metric Histogram of hitting count Average value is not important The more high hitting case, the higher IoT. Var.  worse performance Assumptions –Total 48 DRU –Case 1 : resource loading ratio is 1/6  8 DRU (24 tiles) –Case 2 : resource loading ratio is 2/6  16 DRU (48 tiles)

7. SLS Verification (1) Sequential Sector ID Case 1.Difference is too small. Nothing worth to compare. 7/17

8. Random Sector ID Case 1.A noticeable tendency in cell edge performance Samsung’s is better up to 16.3% SLS Verification (2) 8/17 16.3% 12.3% 7.6% 5.0% 2.7% 0%

9. Theory Goal of Permutation –Interference Averaging, especially not in Full Loading Situation Worse Permutation –Would result in similar logical-physical mapping among neighbor cells Eventually –Higher NI fluctuation results in error rate –This impacts on cell edge performance 9/17

10. NI Fluctuation Observed Metric : Histogram of MS’s NI Variance –MS_NI : NI power of the tones which are assigned to the MS –MS drop and collect total 570 MSs’ MS_NI –Calculate variance along time for every MS  V 1 ~V 570 Time duration : total 2700 UL subframes Total number of drop : 11 drops 10/17 LGE’s has larger Mean{V i } LGE’s has larger Var{V i } High error rate Worse Performance [Note] See Appendix 2 for further cases (e.g. 4/6, 6/6 loading)

11. Hitting Count Verification Methodology –Select 2 Sector IDs from total 768 Total number of cases : combination 2 among 768 = 294528 cases –Count the number of tiles two sectors share Full loading case  obviously 100% hitting Not full loading case  worst case is 100% hitting –Results metric Histogram of hitting count Average value is not important The more high hitting case, the higher IoT. Var.  worse performance Assumptions –Total 48 DRU –Case 1 : resource loading ratio is 1/6  8 DRU (24 tiles) –Case 2 : resource loading ratio is 2/6  16 DRU (48 tiles)

12. Case 2 Hitting Count Results Case 1 –For more cases, open this sheet Number of tiles having collision

13. Downlink Subcarrier Permutation Part 2 13/17

14. Evaluation methodology –User drop on the desired cell, which is located in (radius, theta) # of IDcell for desired cell = 0 Radius is variable, but theta is fixed as 30 degree –Varying parameter of “radius” –Select 7 strongest interferers Calculating only path-loss according to the distance between MS and BSs. Not considering shadowing –Calculate SINR LLS in Multi-cells environment (1) Cell ID Configuration –Increasing sector ID –ISD = 1.5km 1 2 3 67 910 11 13 14 1516 17 18 20 21 23 2425 26 2728 3031 32 3334 35 3637 38 4940 41 4243 44 4546 47 4849 50 5152 53 5455 56 0 5 4 8 19 22 29 12 [ Example] MS is located in (radius, theta) = (0.75*ISD/2, 30) 7 strongest interferers(I 1 ~I 7 ) = 8, 19, 5, 4, 12, 22, 29 SINR=6.36dB MS

15. LLS in Multi-cells environment (2) Simulation conditions –Working scenarios –Number of DRUs / LRUs / Miniband allocation Half loading in DRUs(Ex. # of DRUs = 8, # of LRUs = 4) Half loading in Miniband based CRU –Assuming random QPSK modulated data bursts are transmitted –Assuming random sequence for CRU/DRU allocation sequence –Channel condition: PedB, 3km/h –MIMO configuration: 2x2 SFBC –Pilot Structure Pilot power = 3 dB Interlaced pilot structure Freq. Partition# of subbands# of minibands# of PRUs in FP i Scenario #1 FP0 62448 FP1 ~ FP3 000

16. FER vs SINR (1) Evaluation Results –# of DRUs=4 –# of DRUs=5

17. FER vs SINR (2) Evaluation Results –# of DRUs=6 –# of DRUs=7

18. FER vs SINR (3) Evaluation Results –# of DRUs=8 –# of DRUs=9

19. FER vs SINR (4) Evaluation Results –# of DRUs=10 –# of DRUs=11

20. Conclusion Uplink Tile Permutation –Samsung’s is better in Hitting Count Results –It will cause better interference averaging Downlink Subcarrier Permutation –Similar performance in both sequential and random case –Because BS Tx power is constant For DL and UL, it is natural to be same formula and permutation sequence.

21. Proposed Text for AWD (1)  PermSeq() is the permutation sequence of length L DRU,FPi and is determined by SEED={IDcell*1367} mod 2 10. The permutation sequence is generated by the random sequence generation algorithm specified in Section 15.3.5.3.4. generated by a function or by a lookup table;  g(PermSeq(),s,m,l,t) is a function (TBD) with value from the set [0, L DRU, FPi -1], which is defined as follows.; g(PermSeq(),s,m,l,t) = {PermSeq[{f(m,s)+s+l} mod L DRU,FPi ] +DL_PermBase} mod L DRU,FPi, where DL_PermBase is an integer ranging from 0 to 31(TBD), which is set to preamble IDcell.  f(m,s) = (m+13·s) mod L SP,l. is a function (TBD) with value from the set [0, L SP,l -1]. [Remedy-1: Change the text from line 35 to 38 on the page 31, in 15.3.5.3.3, as follows:] 21/14

22. Proposed Text for AWD (2) 15.3.5.3.4 Random sequence generation The permutation sequence generation algorithm with 10-bit SEED (S n-10, S n-9,…,S n-1 ) shall generate a permutation sequence of size M by the following process: 1)Initialization A.Initialize the variables of the first order polynomial equation with the 10-bit seed, SEED. Set d 1 = floor(SEED/2 5 ) + 1 and d 2 = SEED mod 2 5. B.Initialize the maximum iteration number, N=4. C.Initialize an array A with size M with the numbers 0, 1, …, M-1 (i.e. A[0]=0, A[1]=1, …, A[M-1]=M-1). D.Initialize the counter i to M-1. E.Initialize x to -1. 2)Repeat the following steps if i > 0 A.Initialize the counter j to 0. B.Repetition loop as follows, a.Increment x and j by 1. b.Calculate the output variable of y = {(d 1 *x + d 2 ) mod 1031} mod M. c.Repeat the above step a. and b., if y  i and j<N. C.If y  i, set y = y mod i. D.Swap the i-th and the y-th elements in the array (i.e. perform the steps Temp= A[i], A[i]= A[y], A[y]=Temp). E.Decrement i by 1. 3) PermSeq[i] = A[i], where 0  i<M. [Remedy-2: Insert the text in line 48 on the page 31, in 15.3.5.3.3, as follows:] 22/14

23. Proposed Text for AWD (3) Tile(s,n,t) = TBD L DRU,FPi  n + g(PermSeq(), s, n, t)  PermSeq() is the permutation sequence of length L DRU,FPi and is determined by SEED={IDcell*1367} mod 2 10. The permutation sequence is generated by the random sequence generation algorithm specified in Section 15.3.5.3.4.  g(PermSeq(),s,n,t) is a function of s, n, t and PermSeq(), which is defined as follows: g(PermSeq(),s,n, t) = {PermSeq[(n+107*s+t) mod L DRU,FPi ]+UL_PermBase} mod L DRU,FPi, where UL_PermBase is an integer ranging from 0 to 31(TBD), which is set to preamble IDcell. [Remedy-3: Change the text in line 1 on the page 42, in 15.3.8.3.3, as follows:] 23/14 [Remedy-4: Insert the text in line 13 on the page 42, in 15.3.8.3.3, as follows:]

24. Appendix 1: Parameters for UL SLS 24/#NN ParameterValueParameterValue Carrier frequency (GHz) 2.5 GHz Site to site distance (m) 1500 m System bandwidth (MHz) 11.2 MHz Number of users per sector 10 Reuse factor 1 Channel Ped B, 3km/h 100% Frame ( Preamble +DL +UL ) duration 5 ms (TDD, 29:18) Max power in MS (dBm) 23 dBm Number of OFDM symbols in UL Frame 18 symbols (3 subframes = 6 symbols per subframe) Antenna type 1x2 SIMO FFT size (tone) 1024 HARQ On (Max retrans : 4 / Sync) Useful tone 864 Target IoT Level 10 dB Tile structure 6x6 DRU Link to system mapping RBIR Number of LRU 48 Scheduler type PF Number of tile per LRU 3 tiles PF exponent 1.0 Resource assignment block 8 LRU Penetration loss[dB] 10dB Number of user per subframe 6 user Overhead No control channel, only pilot Power Control Open loop power control UL Target IoT value 10dB

25. Appendix 2: NI Fluctuation Observed Metric : Histogram of MS’s NI Variance 25/17